7644256

Method in Pipelined Data Processing

PublishedJanuary 5, 2010
Assigneenot available in USPTO data we have
Technical Abstract

Patent Claims
7 claims

Legal claims defining the scope of protection, as filed with the USPTO.

1

1. A method in a processor, in which data is processed in a pipelined manner, the data being included in a plurality of contexts, comprising a first context ( 3 ), each context passing a plurality of consecutive stages ( 2 a - 2 f ), in addition to which a plurality of operations is adapted to be executed on the contexts, each operation comprising a plurality of consecutive operation steps and the consecutive operation steps of one operation being executed on a context at least two different consecutive stages ( 2 a - 2 f ), the method comprising: at a first stage ( 2 a ), executing an initial operation step ( 6 a ) of a first operation on the first context ( 3 ), and at a second stage ( 2 b ) that consecutively follows the first stage ( 2 a ), subsequently commencing an execution on the first context of an initial operation step ( 7 a ) of a second operation before an execution on the first context ( 3 ) of a following operation step ( 6 b ) of the first operation is completed, wherein, at each clock cycle of the processor, the first context ( 3 ) is received at one of the stages from the preceding stage, the first context is unconditionally moved to a next stage and a subsequent context of a subsequent operation is received at the first stage ( 2 a ).

2

2. A method according to claim 1 , comprising commencing at the first stage ( 2 a ) an execution of the initial operation step ( 6 a ) of the first operation on a second context before the execution on the first context ( 3 ) of the following operation step ( 6 b ) of the first operation is completed.

3

3. A method according to claim 1 , comprising receiving at the second stage a result (R 6 a ) of an execution of the initial operation step ( 6 a ) of the first operation.

4

4. A method according to claim 1 , whereby at least one of the operation steps of the second operation comprises at least two alternative execution paths, and at least two of the alternative execution paths of the operation step are executed.

5

5. A method according to claim 4 , further comprising: obtaining results (R 7 b 1 , R 7 b 2 ) of at least two of the executions of the alternative execution paths, and determining, based on a result (R 6 ) of an execution of an operation step of an operation initiated before the initiation of the second operation, which one of the results (R 7 b 1 , R 7 b 2 ), of the executions of the alternative execution paths, an execution of an operation step of the second operation, following said operation step comprising at least two alternative execution paths, is to be based on.

6

6. A method according to claim 1 , whereby the processor is arranged so that the following operation step ( 6 b ) of the first operation is presented to a programmer as being executed at the first stage ( 2 a ).

7

7. A method according to claim 1 , wherein the first operation comprises a partial operation of executing ( 6 c 1 ) an instruction and a partial operation of writing ( 6 c 2 ) a result of the said instruction execution into a destination in a register, and the second operation comprises the partial operation of fetching ( 7 a 2 I, 7 a 22 ) an operand, the method comprising (a) determining if a position in the register, from which the operand is to be fetched ( 7 a 2 1 , 7 a 22 ) in the second operation, is identical with the destination of the partial operation, of the first operation, of writing ( 6 c 2 ) a result, (b) if the result of the determination in step (a) is negative, fetching ( 7 a 21 ) the operand from the register, and (c) if the result of the determination in step (a) is positive, fetching ( 7 a 22 ) the result of the said instruction execution.

Patent Metadata

Filing Date

Unknown

Publication Date

January 5, 2010

Inventors

Gunnar Nordmark
Thomas Boden

Want to explore more patents?

Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.

Citation & reuse

Analysis on this page is generated by Patentable — an AI-powered patent intelligence platform. AI-generated summaries, explanations, and analysis may be reused with attribution and a visible link back to the canonical URL below. Patent abstracts and claims are USPTO public domain.

Cite as: Patentable. “METHOD IN PIPELINED DATA PROCESSING” (7644256). https://patentable.app/patents/7644256

© 2026 Patentable. All rights reserved.

Patentable is a research and drafting-assistant tool, not a law firm, and does not provide legal advice. Documents we generate are drafts for review by a licensed patent attorney.